OR I G I N A L A R T I C L E

Avoiding chaotic use of social media before, during, and afteremergencies: Design and evaluation of citizens’ guidelines

Marc‐André Kaufhold1 | Alexis Gizikis2 | Christian Reuter3 | Matthias Habdank4 |

Margarita Grinko1

1Institute for Information Systems,

University of Siegen, Siegen, Germany

2European Emergency Number Association

(EENA), Brussels, Belgium

3Science and Technology for Peace and

Security (PEASEC), Technische Universität

Darmstadt, Darmstadt, Germany

4Computer Application and Integration in

Design and Planning (CIK), University of

Paderborn, Paderborn, Germany

Correspondence

Marc‐André Kaufhold, Institute for

Information Systems, University of Siegen,

Siegen, Germany

Email: marc.kaufhold@uni-siegen.de

Funding information

European Union, Grant/Award Number: FP7

No. 608352; German Federal Ministry of

Education and Research, Grant/Award

Number: BMBF No. 13N14351

Abstract

Social media have been established in many natural disasters or human‐inducedcrises and emergencies. Nowadays, authorities, such as emergency services, and citi-

zens engage with social media in different phases of the emergency management

cycle. However, as research in crisis informatics highlights, one remaining issue con-

stitutes the chaotic use of social media by citizens during emergencies, which has

the potential to increase the complexity of tasks, uncertainty, and pressure for

emergency services. To counter these risks, besides implementing supportive tech-

nology, social media guidelines may help putting artefact and theoretical contribu-

tions into practical use for authorities and citizens. This paper presents the design

and evaluation (with 1,024 participants) of citizens’ guidelines for using social media

before, during, and after emergencies.

K E YWORD S

emergency management, guidelines for citizens, representative survey, social media

1 | INTRODUCTION

Social media are a part of everyday life and are also prevalent during

crises, disasters, and emergencies. Reuter and Kaufhold (2018) sum-

marized 15 years of social media use during natural disasters and

human‐induced crises, identifying different usage, role, and percep-

tion patterns between administrative and public stakeholders both in

the real and in virtual realms. New opportunities have emerged for

emergency services to alert and warn of crises (Brynielsson et al.,

2017) and for social media analytics (Stieglitz, Mirbabaie, Ross, &

Neuberger, 2018). Palen and Hughes (2018) conclude that social

media also promote socio‐technical innovations, such as citizen

reporting, community‐oriented computing, collective intelligence and

distributed problem solving, and digital volunteerism. However, one

remaining issue is that citizens’ activities coordinated via social media

have the potential to increase the complexity of tasks, uncertainty,

and pressure for emergency services, for instance, if volunteers them-

selves are endangered (Perng et al., 2012). Given the risks of chaotic

social media use (Kaewkitipong, Chen, & Ractham, 2012), guidelines

may help authorities implement social media into their organizational

culture (Reuter, Ludwig, Kaufhold, & Spielhofer, 2016) and encourage

citizens to foster good practice and prevent misuse of social media

during emergencies (Reuter & Spielhofer, 2017).

While there are plenty of guidelines for using social media in

general or during emergencies from the perspective of organizations,

few guidelines approach the citizens’ perspective and none of them

has been evaluated with citizens. Thus, our study aims to contribute

to the design and evaluation of citizens’ guidelines for using social

media in emergencies, while the latter aspect of evaluation is the

main focus of this paper. The aim of the guidelines is to approach

various issues that occur with social media usage from citizens’ per-spective during crises and that are not adequately addressed by

existing guidelines. Such issues include task complexity and uncer-

tainty, information overload, the need for timely and trustworthy

information, and the effective organization of volunteer groups and

DOI: 10.1111/1468-5973.12249

J Contingencies and Crisis Management. 2018;1–16. wileyonlinelibrary.com/journal/jccm © 2018 John Wiley & Sons Ltd | 1

help for emergency services. We will outline them in detail in the

next section. Our main research question is: How do citizens per-

ceive guidelines for social media use in emergencies (RQ1)? To guide

the analysis of our results, we formulated further five sub‐questions:

� To what extent do participants agree to the presented guidelines

on social media behaviour before, during, and after emergencies

(RQ1.1)?� How consistent is the agreement with guidelines across the cate-

gories, that is, concerning the use of social media before, during,

and after emergencies (RQ1.2)?� To what extent is the attitude towards the guidelines affected by

smartphone and social media usage in daily life and in emergen-

cies (RQ1.3)?� How do demographic factors influence the agreement with

guidelines concerning the use of social media in emergencies

(RQ1.4)?� What are caveats and improvement suggestions towards the

guidelines on social media in emergencies (RQ 1.5)?

Firstly, related work on the use and challenges of social media in

emergencies and an analysis of existing guidelines for the use of

social media will be discussed (Section 2). Based on this, we

designed citizens’ guidelines for evaluation (Section 3). To gather

quantitative and qualitative feedback on our research questions, we

then conducted a representative online survey in Germany in July

2017 (Section 4). The paper concludes with a short summary of

results, discussion, and outlook for future research (Section 5).

2 | RELATED WORK: CHALLENGES OFSOCIAL MEDIA IN EMERGENCIES ANDEXISTING GUIDELINES

Based on the potentials of social media in the emergency manage-

ment cycle (EMC), comprising the phases of mitigation, prevention,

response, and recovery (Baird, 2010), we present and discuss exist-

ing guidelines for the use of social media. In this paper, we use the

term emergency as a state of collective disruption, caused by natural

hazards (i.e., earthquakes, floods, hurricanes, and tsunamis) or

human‐induced disasters (i.e., accidents, attacks, shootings, terrorism,

and uprisings) that are of a sudden and unexpected nature. There-

fore, they threaten societal values and structures, and require

authorities to make decisions during uncertain circumstances (Boin,

Stern, & Sundelius, 2005).

2.1 | Use and challenges of social media inemergencies

For 15 years, the public has used social media in emergencies (Reu-

ter & Kaufhold, 2018), with 2012 Hurricane Sandy being one of the

most prominent cases (Caragea, Squicciarini, Stehle, Neppalli, &

Tapia, 2014; Homeland Security, 2013; Kogan, Palen, & Anderson,

2015) but also the 2008 Sichuan earthquake (Li & Rao, 2010; Qu,

Wu, & Wang, 2009) or 2013 European floods (Kaufhold & Reuter,

2016; Reuter & Schröter, 2015). Accordingly, a mixed‐method study

with authorities and citizens outlines that, in times of crisis, Face-

book is perceived as a channel to reach the general public and

where people debate and connect with family and friends, whereas

Twitter is perceived as an elite channel which primarily serves as an

early warning channel (Eriksson & Olsson, 2016). Research outlines a

variety of potentials of social media use, that is, increasing the resili-

ence of citizens (Jurgens & Helsloot, 2017). Panic, looting, and help-

lessness or dependency on external rescuers in crisis situations are

rare, contrary to popular belief (Helsloot & Ruitenberg, 2004).

Instead, self‐organization and help are a common reaction: Citizens

of affected areas assist emergency services by social aftercare for

victims and relatives online (Mirbabaie & Zapatka, 2017), search for

missing people by sharing and using information on microblogging

platforms (Qu, Huang, Zhang, & Zhang, 2011), and coordinate and

mobilize volunteers through social media (Albris, 2017; Kaufhold &

Reuter, 2016).

Despite these potentials, research also outlines a variety of risks

associated with the use of social media (Hiltz & Plotnick, 2013; Plot-

nick & Hiltz, 2016; Reuter et al., 2016). A study on the 2011 Nor-

way attacks revealed that citizens’ activities coordinated via social

media increased the complexity of tasks, uncertainty, and pressure

for emergency services, since volunteers endangered themselves by

autonomous rescue attempts (Perng et al., 2012). Likewise, Kaewki-

tipong et al. (2012) report on the risks of chaotic social media use,

including “information redundancy, information inconsistency, chaos

and rumors among all social communities” (p. 13), that occurred dur-

ing the 2011 Thailand flooding disaster. A further study on the 2010

Haiti earthquake highlights that uncertainty can be generated by

unreliable information and mistakes due to chaotic and disorganized

work of volunteers (Valecha, Oh, & Rao, 2013). During the 2013

floods in Dresden, Facebook groups such as “Fluthilfe Dresden”(Flood Help Dresden) distributed and spread out citizen volunteers

successfully, that is, for filling and piling sandbags, although in some

cases more people were mobilized than the response effort could

accommodate (Albris, 2017). Thus, the convergence of volunteers

led to suboptimal outcomes and authorities feared dilettante beha-

viour by citizen volunteers, indicating the need for guidance to

improve citizens’ behaviour both in real and in virtual spaces (Reuter,

Heger, & Pipek, 2013). Accordingly, Schmidt, Wolbers, Ferguson, and

Boersma (2017) outline the challenges of governing self‐organizationof citizens, connecting online platforms with on‐site response initia-

tives, fostering inclusiveness, and managing information.

2.2 | Existing guidelines for the use of social media

Our search revealed that there are many existing guidelines concern-

ing the use of social media in general and for use in emergency situ-

ations (see Section 9). We found no commonly accepted definition

for social media guidelines in emergencies, but multiple aspects in

terms of their applicability, domains, and goals are highlighted in dif-

ferent definitions. For instance, social media guidelines for emergen-

cies intend to “enable citizens using new mobile and online

technologies to actively participate in the response effort” (Belfo et

2 | KAUFHOLD ET AL.

al., 2015) and to “enhance the safety and security of citizens by sup-

porting both citizens, and public authorities, in their use of social

media to complement their crisis management efforts” (Helsloot et

al., 2015). Table 1 contains guidelines and handbooks from different

institutional backgrounds, ranging from civil institutions, cities, and

authorities to the military. In addition, organizations, companies, and

authorities publish guidelines for the use of social media by their

employees. These guidelines usually focus on the employee's private

use of social media and rules regarding the company or authority,

therefore lacking the use perspective of the citizens as public actors.

The guidelines mostly point to law, behaviour, data security issues,

and non‐disclosure obligations. Yet, these handbooks do not concern

social media usage with regard to the EMC.

As Table 2 shows, there are guidelines for the specific use of social

media in emergencies. From our analysed set, only the guidelines of

the European projects COSMIC and iSAR+ as well as the Emergency

2.0 Wiki cover the citizens’ perspective. The COSMIC guidelines pro-

vide recommendations in terms of preparation, seeking aid and infor-

mation, providing aid, mobilization, as well as recording and sharing

(Helsloot et al., 2015). Over 14 pages, the different advices are illus-

trated by text, examples, pictures, and bullet points. In these guideli-

nes, only the authorities’ recommendations (represented by

parenthetical crosses in Table 2) but not the citizens’ guidelines refer

to the EMC. In contrast, the iSAR+ guidelines provide three pages of

“recommendations and explanations on how to prepare for a crisis sit-

uation, what to consider when the incident is about to happen, ongo-

ing, and when it is over” (Belfo et al., 2015). The guidelines are divided

into the phases of preparation, warning, acute incident, and aftermath,

giving a short explanation and several bullet points with concrete

advice for each. Similarly, the Emergency 2.0 Wiki (2015) provides

“tips for the public” on how to behave before, during, and after emer-

gencies, including a general introduction to social media, advice on

finding and sharing information during the different phases, communi-

cating with family and friends, or saving battery power. Both the

iSAR+ guidelines and Emergency 2.0 Wiki organize their recommenda-

tions based on the EMC, although the mitigation phase is not included.

These guidelines mostly use bullet points and pictures for advice and

illustration, and they are distributed over different wiki pages.

2.3 | Research gap and objective

Despite the potentials of citizens’ use of social media before, during,

or after emergencies, their activities potentially increase the com-

plexity of tasks, uncertainty, and pressure for emergency services

(Perng et al., 2012), for instance, by irrelevant and inconsistent infor-

mation, information overload, and mistakes due to chaotic and disor-

ganized work of volunteers (Kaewkitipong et al., 2012; Valecha

et al., 2013). Thus, guidelines may motivate citizens to foster good

practice and prevent misuse of social media during emergencies

(Reuter & Spielhofer, 2017). Previous research indicates that a vari-

ety of general or emergency‐specific guidelines exist, but most of

them focus on the authorities’ or organizations’ perspective of social

media use. Only three guidelines explicitly contained recommenda-

tions for citizens before, during, and after emergencies, varying in

length and media richness (Belfo et al., 2015; Emergency 2.0 Wiki,

2015; Helsloot et al., 2015). Considering feedback from

TABLE 1 List of guidelines for the use of social media in general

Title Publisher

Verification Handbook—An ultimate guideline on digital age sourcing for

emergency coverage (2014)

European Journalism Centre, EU

Social engagement handbook 2.0 (2012) American Red Cross, US

Social media handbook (2014) United States Army, US

Ein Leitfaden zum Umgang mit Social Media im DRK (2012) Deutsches Rotes Kreuz (German Red Cross, DRK), DE

Social‐Media‐Guideline—Empfehlungen für einen sicheren Umgang mit

sozialen Medien (2012)

Berliner Feuerwehr (Berlin Fire Department), DE

Verhalten in sozialen Netzwerken (2011) Bundesanstalt Technisches Hilfswerk (Federal Agency for Technical

Relief), DE

Das soziale ins Netz bringen—die Caritas und soziale Medien (2014) Deutscher Caritasverband (German Caritas Association), DE

Social media guidelines for IFRC staff (2009) International Federation of Red Cross and Red Crescent Societies (IFRC)

Social Media—Guidelines for Canadian Red Cross Staff and

Volunteers (2013)

Canadian Red Cross, Canada

Rotkreuz‐Social‐Media‐Policy (2010) Österreichisches Rotes Kreuz (Austrian Red Cross), Austria

ACT Government Social Media Policy Guidelines (2012) ACT Government, Australia

Social Media in der Hamburgischen Verwaltung—Hinweise,

Rahmenbedingungen und Beispiele (2011)

Freie Hansestadt Hamburg (Free Hanseatic City of Hamburg), DE

Social Media Guidelines and Best Practices—Facebook (2012) Centers for Disease Control and Prevention (CDC), US

Social Media Guidelines and Best Practices—CDC Twitter Profiles (2011)

The health communicator's social media toolkit (2011)

CDC's Guide to Writing for Social Media (2012)

KAUFHOLD ET AL. | 3

practitioners’ workshops (Section 3), these guidelines were perceived

as too extensive for reaching a broad audience, and, to our best

knowledge, none of them was evaluated with citizens in terms of

their practical value. Thus, besides presenting the design of concise

citizens’ guidelines that are suitable for distribution to large audi-

ences before, during, and after emergencies, the main aim of this

paper is the quantitative and qualitative evaluation of these guideli-

nes according to our research questions.

3 | METHODOLOGY

This section describes the design and evaluation methodology of the

citizens’ guidelines for using social media before, during, and after

emergencies.

3.1 | Design of citizens’ guidelines

The citizens’ guidelines were designed to convey general aspects

while using social media and information on how to behave before,

during, and after an emergency. Firstly, we conducted literature

reviews on the potentials and challenges of social media from the per-

spective of citizens and then reviewed guidelines on the use of social

media in general and during emergencies. The guidelines listed in

Tables 1 and 2 were identified by searching in Google, scanning

through related European projects on social media, and considering

expert recommendations of an end‐user advisory board (EAB) of our

European project. The search focused on English guidelines, although

the project consortium recommended a few German guidelines that

were considered subsequently. Although all found guidelines were

considered, only three contained recommendations for citizens in

emergencies and thus were relevant for the analysis in this paper. For

the design of our guidelines, no extensive comparative analysis was

performed between these three guidelines, but their content was pre-

sented as input during an EAB workshop, comprising 18 participants

from emergency services (O'Brien et al., 2016). Based on these guide-

lines, EAB members and practitioners prioritized the most important

information that should be communicated to citizens with the inten-

tion to keep the guidelines concise. Under consideration of their input,

further literature, and (revisiting of) guidelines, a first draft of the citi-

zens’ guidelines was created. The draft was presented at another EAB

workshop, comprising 15 participants emergency services, to gather

insights for the design of the final guideline versions (Gizikis, Susaeta,

et al., 2017). Since the focus of this paper lies within the evaluation

but not the design of the guidelines, please find additional information

TABLE 2 List of guidelines for the use of social media in emergencies

Title Publisher ES C EMC

Guidelines for the use of new media in crisis situations (2015) COSMIC project, EU x x (x)

Warning and informing Scotland using social media in

emergencies (2012)

Scottish Government, Scotland x

Social media for emergency management—a good practice

guide (2014)

Wellington Region Emergency Management Office,

New Zealand

x (x)

Emergency 2.0 Wiki (2015) Emergency 2.0 Wiki x x x

Social Media in an emergency: Developing a Best Practice

Guide Literature Review (2012)

Opus International Consultants Limited, New Zealand x (x)

iSAR+ Guidelines (2015) iSAR+ Project, EU x x x

Using social media for emergency notifications—7 questions

for emergency managers to consider

Twenty First Century Communications, Inc., US x

Social Media in Emergencies—UNICEF Guidelines for

Communication and Public Advocacy (2012)

United Nations Children's Fund (UNICEF) x

Smart tips for category 1 responders using social media in

emergency management (2012)

Defense Science and Technology Laboratory

(DSTL), UK

x (x)

Crisis communications and social media—A best practice

guide to communicating an emergency (2014)

International Air Transport Association (IATA) x

Next Steps: Social Media for Emergency Response (2012) Homeland Security, US x

Using Web 2.0 applications and Semantic Technologies to

strengthen public resilience to disasters (2013)

Disaster 2.0, EU x

Bevölkerungsschutz: Social Media (2014) Bundesamt für Bevölkerungsschutz und

Katastrophenhilfe (Federal Office of Civic Protection

and Disaster Assistance, BBK), DE

x

The Use of Social Media in Risk and Crisis

Communication (2014)

Organization for Economic Co‐operation and

Development (OECD)

x

Leitfaden Krisenkommunikation (2014) Bundesministerium des Innern (Federal Ministry of

the Interior, BMI), DE

x

ES: emergency service guidelines; C: guidelines for citizens; EMC: emergency management cycle.

4 | KAUFHOLD ET AL.

on a related project deliverable (Gizikis, O'Brien, et al., 2017). The

resulting guidelines are presented in Table 3.

3.2 | Evaluation of the citizens’ guidelines

We conducted a representative online survey (N = 1024) of the adult

German population in July 2017, using the ISO‐certified panel provi-

der GapFish (Berlin). They guarantee panel quality, data quality, and

security, as well as survey quality through various (segmentation)

measurements for each survey within their panel of 180,000 active

participants. Our overall survey covered also other topics from our

research project. In this work, we are examining five of these (Q22–26), related to social media guidelines. Firstly, participants were asked

closed questions on the perceived importance of the guidelines

before (Q22), during (Q23), and after (Q24) an emergency (“How do

you rate the following recommendations for usage of social media

[on/during/after] an emergency?”), and general aspects while using

social media (Q25: “How do you rate the following, general aspects

for using social media in an emergency?”). The responses were mea-

sured using a five‐point Likert scale ranging from “very important” to

“not important at all.” Secondly, we asked for open‐ended criticism or

improvement suggestions for the presented guidelines about the use

of social media before, during, and after a crisis (Q26).

3.2.1 | Characteristics of survey participants

The sample of survey respondents (N = 1024) was adapted to the dis-

tribution of age, gender, region, education, and income according to the

general German population (German Federal Agency for Civic Educa-

tion, 2016; Statista, 2016; Statistisches Bundesamt (Destatis), 2016).

Preselection of participants was carried out by the external survey

panel provider (GapFish) so that we received the already adapted sam-

ple. Based on these statistics, our sample consisted of 49.5% female

(n = 507) and 50.5% male (n = 517) respondents. They were between

18 and 64 years old, nearly half of them being 45 and older (n = 492,

48%). We recruited participants from every federal state of Germany,

TABLE 3 Social media guidelines

General

Aspects

� Interact with respect and courtesy (Bundesanstalt Technisches Hilfswerk, 2011)� You are responsible for your writing, think of possible consequences (Bundesanstalt Technisches Hilfswerk, 2011)� Protect your privacy and check the privacy settings (Deutsches Rotes Kreuz, 2012)� Respect intellectual property rights, including pictures, graphics, audio and video files (Berliner Feuerwehr, 2012).Add references, make quotations marked (Deutscher Caritasverband, 2014), specify sources, and without approval, donot talk about a third person (Daimler, 2012)� Verify your information before posting (Helsloot et al., 2015)� Correct a mistake if you made one (Kaewkitipong et al., 2012)

Preparation Be prepared� Know the social media accounts of your local and national ES and follow them. This will help find real‐time information

during an emergency (Helsloot et al., 2015)� Read what to expect from ES in social media. Are they always online? Do they reply to posts in social media (Reuteret al., 2016)?� Look for apps that ES provide and download them to stay informed during an emergency (Reuter et al., 2017)� Follow the information from ES on how to prevent and stay safe during emergencies (Helsloot et al., 2015)

Response Stay up‐to‐date� Follow official accounts and local organizations to get information updates (Helsloot et al., 2015)Social media does not replace 112� Remember you can use social media for information updates, but it does not replace emergency calls (Helsloot et al.,

2015). If in danger, always call 112 firstBe responsible and avoid spreading rumours!When you post information about an emergency in social media:� Always mention the ES account or include any already used hashtags. When possible report a location and use photos

(Helsloot et al., 2015)� Tell only facts and don't send information you are not certain about (Helsloot et al., 2015)� Share only official and reliable information and avoid spreading rumours! The spreading of false information can threatenthe smooth deployment of rescue teams and put you and your relatives at additional risk (Perng et al., 2012)� If you spot or shared false information, please correct it (Helsloot et al., 2015)� Forward received official messages to your contacts or share them (Helsloot et al., 2015)

Volunteering initiatives� Look for emergent volunteer initiatives in Facebook groups, Google crisis maps, or trusted users in Twitter; they may helpto increase the impact of your activities (Kaufhold & Reuter, 2016)!� If you intend to initiate your emergent volunteer initiative, please check for existing initiatives first and carefully chose thescope of your possible contribution

Recovery � Follow official accounts and local organizations to get information updates (Helsloot et al., 2015). Communicate even aftera crisis and use social media for the processing of the event� Give feedback to the authorities (Reuter et al., 2016)� Restore missing contact and ask for the welfare of family and friends (Qu et al., 2011)� Help others reconstructing/handling the event (Mirbabaie & Zapatka, 2017)

KAUFHOLD ET AL. | 5

where the largest sample came from North Rhine‐Westphalia (n = 228,

22%) and Bavaria (n = 165, 16%). Only 1% (n = 9) of participants did

not graduate from a school, while 15% (n = 160) held a degree from a

university or college. Over two‐thirds (n = 707, 69%) earned between

1.500€ and 3.500€. Concerning the smartphone usage, almost half of

respondents indicated to use it daily (n = 500, 49%). A similar result

could be found for daily usage of social media, namely Facebook

(n = 475, 46%), instant messaging services (n = 449, 43%), and You-

Tube (n = 301, 29%). Another third stated an hourly usage of smart-

phones (n = 441, 43%) and instant messengers (n = 339, 33%), while

19% (n = 196) claimed to actively post in social media daily and an

almost equal amount (n = 197) at least once a week. As for behaviour

in crisis situations, over half of participants (n = 556, 54%) indicated to

have used social media in such an event. Slightly more (n = 607, 59%)

expected to see messages by critical infrastructure (CI) operators on

social media in the case of an infrastructure failure.

3.2.2 | Quantitative and qualitative analysis

For the quantitative analysis, the survey data were extracted and

analysed using IBM SPSS Statistics 23, a software package for ana-

lysing quantitative data (IBM, 2014). Microsoft Excel was used for

qualitative coding and for the design of the diagrams showing rela-

tive amounts of participant responses (see Section 4). The analysis

consisted of three key steps: (1) preparing the data including assign-

ment of missing data values and combination of categories of demo-

graphic background variables. This occurred by excluding participants

where an answer to a closed question was missing, which would

have made quantitative analysis unreliable, thus reducing the number

of participants from N = 1,069 to N = 1,024. (2) Exploring basic fre-

quencies for each question. (3) Using cross‐tabulations with chi‐squared tests to explore any significant differences across different

types of respondents in relation to gender, age, region, income, and

education level, as well as smartphone and social media usage. (4)

Determining correlations between ordinal items using Spearman's Rho.

The qualitative analysis of our open‐ended survey question was

based on the inductive approach of grounded theory (Strauss, 1987).

We used open coding to derive categories from the more qualitative

open‐ended responses by carefully reading and aggregating categories.

In a first iteration, to achieve a quick overview of the interesting and

relevant topics, two coders went through the open‐ended responses

independently and proposed a set of preliminary categories. The cate-

gories were discussed in tandem to agree upon a shared set of cate-

gory codes. The previously acquired knowledge from the literature

review and quantitative analysis was used to increase theoretical sen-

sitivity. In a second iteration, the coders independently assigned each

open‐ended response to one or multiple categories. Finally, the codes

were jointly checked to reach an agreement between the coders. Each

quotation is referenced with the participant's response identifier.

4 | RESULTS

This section presents the results of the quantitative and qualitative

evaluation of citizens’ attitudes towards the guidelines on social

media use before, during, and after emergencies.

4.1 | Importance of individual recommendations ofthe guidelines

To what extent do participants agree to the presented guidelines on

social media behaviour before, during, and after emergencies

(RQ1.1)? Before an emergency (see Figure 1), interestingly, more

than two‐thirds highlighted the importance of knowing the social

media accounts of local and national emergency services (ES) (70%)

or following their information on how to prevent and stay safe dur-

ing emergencies (66%). In contrast, only about half think it is essen-

tial to look for apps that emergency services provide and download

them to stay informed during an emergency (52%). Even fewer par-

ticipants agree to read what to expect from emergency services in

social media (45%).

The guideline which participants consider as most important to

follow during an emergency (see Figure 2) is the fact that social

media do not replace the emergency call and should rather be used

for information updates (82%). Seventy‐nine per cent of the

2% 2% 3% 4%4% 4%10% 10%

25% 28%

35%41%

35%36%

32%30%

35%30%

20% 15%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Know the social media accounts of your local and na�onal ES

Follow the informa�on from the ES on how to prevent and stay

safe

Look for ES apps to stay informed during an emergency

Read what to expect from ES in social media

not important at all rather unimportant neutral rather important very important

F IGURE 1 How do you rate thefollowing recommendations for usage ofsocial media before an emergency? (Q22)

6 | KAUFHOLD ET AL.

participants think it is essential to share only official and reliable

information and avoid spreading rumours. Moreover, three quarters

agree it is vital to tell only facts and to not send unreliable infor-

mation. Seventy‐four per cent rate it as important to correct false

information if they spot or share it, and 67% of the participants

want to always stay up‐to‐date. Also, 66% find it essential to

always mention the emergency service account or include any

already used hashtags and forward received official messages to

contacts or share them. More than a half of the respondents (55%)

assume that checking for existing initiatives if they intend to initi-

ate an emergent volunteer initiative is important. It is least, but still

important for the respondents (47%) to look for emergent volun-

teer initiatives in Facebook groups, Google crisis maps, or trusted

users on Twitter.

The question about the importance of different activities after

an emergency (see Figure 3) received the most negative responses.

However, about half of the participants still assume it is essential

to give feedback to the authorities (54%). Also, over half of the

respondents find it important to follow the official accounts and

local organizations to get information updates and to communicate

even after a crisis, as well as use social media for the processing of

the event (57%). It is even more important for the participants to

help others to reconstruct and handle the event, and to restore

missing contacts and ask for the welfare of family and friends (61%

each).

Besides the importance of guidelines during specific phases of an

emergency, in terms of general aspects while using social media (see

Figure 4), about four out of five of the participants find it very or

1% 2% 1% 2% 2% 3% 2% 3% 4%2% 2% 3% 3% 4% 4% 4% 6% 8%15% 17% 20% 21%27% 28% 28%

37%41%

19%22% 20%

27%

37% 33% 33%

28%

28%63% 57% 55%47%

30% 33% 33%27%

19%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Social media does not replace

emergency calls.

Share only official and

reliable informa�on

Tell only facts and don't

send uncertain

informa�on

If you spot or shared false informa�on,

please correct it

Always stay up-to-date

Men�on the ES account,

include hashtags,

loca�ons and photos

Forward received official

messages to your contacts

Check for exis�ng

volunteer ini�a�ves

Look for emergent volunteer in�a�ves

very unimportant rather unimportant neutral rather important very important

F IGURE 2 How do you rate the following recommendations for usage of social media during an emergency? (Q23)

3% 2% 3% 3%5% 6% 7% 9%

32% 30% 33% 34%

33% 34% 32% 33%

28% 27% 25% 21%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Help others reconstruc�ng/ handling the event

Restore missing contacts and ask for welfare of family and

friends

Follow the official accounts and con�nue communica�on

Give feedback to the authori�es

not important at all rather unimportant neutral rather important very important

F IGURE 3 How do you rate thefollowing recommendations for usage ofsocial media after an emergency? (Q24)

KAUFHOLD ET AL. | 7

rather important to correct a mistake if they made one (83%), and to

know it is important that they are responsible for their writing and

that they should remember possible consequences (82%). Addition-

ally, they agree it is essential to interact with respect and courtesy

(82%) and to verify information before posting (81%). Participants

also find it important to protect their privacy and check privacy set-

tings (79%), as well as to respect intellectual property rights (75%).

How consistent is the agreement with guidelines across the cate-

gories, that is, concerning the use of social media before, during, and

after emergencies (RQ1.2)? Taking a closer look directly at the ques-

tionnaire items concerning guidelines, we revealed a significant cor-

relation (Spearman's Rho) between the responses for all guidelines

(p < 0.0001), the values ranging between r = 0.131 and r = 0.785.

For all correlations between the responses, see Table A1 in the

Appendix 2. In the following, we will highlight a few significant find-

ings to explore the research question mentioned above.

If we examine correlations within groups, we observe especially

high significances over r = 0.600. The highest correlation coefficients

were found between items of Q25 (general social media guidelines).

Those who agreed to follow information provided by ES tended to

agree with looking for ES apps to a higher degree (r = 0.610,

p < 0.0001). There is also particularly high correlation of the guideline

“tell only facts” with “share only official and reliable information”(r = 0.778, p < 0.0001). Furthermore, attitude towards the two guide-

lines on helping others and restoring contacts were especially highly

correlated (r = 0.718, p < 0.0001). Opinions on verification and cor-

rection of own posts also showed a significant correlation (r = 0.785,

p < 0.0001), as do respect, courtesy, and responsibility for own writ-

ing (r = 0.740, p < 0.0001).

To reveal the relationship between the individual categories we

addressed in the questions, we also carried out Spearman correla-

tions between the clustered items of each group (see Table 4). All

groups were significantly related and showed a high correlation coef-

ficient, mostly greater than r = 0.600. Only for Q25 (general guideli-

nes), the relationship to guidelines before and after an emergency as

slightly lower than r = 0.500.

4.2 | Influencing factors on the attitude towardsthe guidelines

Via chi‐squared tests, we examined whether demographic factors as

well as smartphone and social media usage habits, both in daily life

1% 1% 1% 2% 1% 2%2% 1% 2% 2% 3% 5%

15% 16% 15% 16% 16%18%

23% 25% 23% 24% 26%25%

60% 57% 58% 57% 53% 50%

0%

10%

20%

30%

40%

50%

60%

70%

80%

90%

100%

Correct a mistake if you made one

You are responsible for your wri�ng

Interact with respect and courtesy

Verify your informa�on before

pos�ng

Protect your privacy and check the privacy se�ng

Respect intellectual property rights

not important at all rather unimportant neutral rather important very important

F IGURE 4 How do you rate the following, general aspects for using social media in an emergency? (Q25)

TABLE 4 Correlations between responses for the individual questions

Q22: Before Q23: During Q24: After Q25: General

Q22: Before r = 0.657, p < 0.001 r = 0.632, p < 0.001 r = 0.425, p < 0.001

Q23: During r = 0.657, p < 0.001 r = 0.690, p < 0.001 r = 0.634, p < 0.001

Q24: After r = 0.632, p < 0.001 r = 0.690, p < 0.001 r = 0.481, p < 0.001

Q25: General r = 0.425, p < 0.001 r = 0.634, p < 0.001 r = 0.481, p < 0.001

8 | KAUFHOLD ET AL.

and in crisis situations, were related to the responses. The factors

have been derived from other questions included in our survey at an

earlier point, where we asked for age and gender, education status

and income, frequency of smartphone usage, social media usage and

publishing posts therein, and familiarity with crisis apps. Similarly, we

also considered the degree to which participants expected critical

infrastructure operators to use social media in case an infrastructure

failure occurs. To provide an overview of these influences, we clus-

tered the response values for each question and carried out chi‐squared cross‐tabulations for the clustered values. Additionally, we

used Spearman correlations to determine the direction of interesting

correlations for a more detailed insight into trends. In Table 5, all

chi‐squared test results and correlations are presented. The depen-

dencies between the demographic variables themselves have not

been calculated.

To what extent is the attitude towards the guidelines affected

by smartphone and social media usage in daily life and in emer-

gencies (RQ1.3)? A higher smartphone usage influenced responses

to questions Q23 (r = 0.083), Q24 (r = 0.059), and Q25

(r = 0.019), leaving out guidelines before an emergency. At the

same time, the more the participants indicate to post in social

media, the more they tend to agree to the statements in ques-

tions Q22 (r = 0.212), Q23 (r = 0.151), and Q24 (r = 0.169),

namely guidelines before, during, and after an emergency. Simi-

larly, for participants’ usage of social media in crisis situations and

CI operator communication expectations, we found significant

effects for all questions:

The more the social media are used in a crisis, the more the par-

ticipants agree to the proposed guidelines on it (r = 0.252; r = 0.162;

r = 0.186, r = 0.010). The same is true for expectations from CI

operators to use social media (r = 0.302; r = 0.282; r = 0.284;

r = 0.202). In contrast to this, general social media and crisis app

usage do not significantly influence the responses.

How do demographic factors influence the agreement with guide-

lines concerning the use of social media in emergencies (RQ1.4)? Gen-

der and age only have a significant influence on the items in question

Q25 (general aspects for the use of social media). Here, a Spearman

correlation indicates that women tended to agree to the general guide-

lines to a higher extent as compared to men (r = 0.132). The same

applies to older participants as compared to younger ones (r = 0.149).

We also found that the higher the income, the more participants

tend to agree to statements in the question on guidelines before an

emergency, Q22 (r = 0.030). In contrast to this, respondents with

lower education show a greater tendency to agree with all state-

ments on guidelines before (r = −0.035) and after an emergency

(r = −0.007), while the opposite was true for general guidelines

(r = 0.039) and those to be adhered to during an emergency

(r = 0.052). However, while chi‐squared tests show a significant

influence of demographic factor, we could not determine the same

for Spearman correlations, which means we cannot make a solid

statement on a trend in this case. Region is the only demographic

factor which does not have a significant influence on any of the

questions.

4.3 | Citizens’ attitudes and suggestions concerningthe guidelines

What are caveats and improvement suggestions towards the guideli-

nes on social media in emergencies (RQ1.5)? In Q25, the last item

asked the participants about their overall opinion on the guidelines:

“The presented guidelines are helpful.” The participants generally

agreed that the guidelines are generally appropriate and helpful, as

72% of participants gave a positive answer. However, the least

amount of participants agreed with this item (37% compared to

more than 50% of all other items in this question). Twenty‐four percent were neutral, and only two per cent each disagreed. To analyse

the reasons for and opinions behind these responses, we further

asked for open comments and present the qualitative results in the

following.

The answers to our open‐ended question Q26 “Which criticism

or improvement suggestions concerning the presented guideline

about the use of social media before, during, and after a crisis do

you have?”, responses were classified into ten meta‐codes: (1) fakenews, (2) gazers, (3) no idea, (4) no answer, (5) no criticism, (6) presen-

tation, (7) scepticism, (8) extent, (9) unnecessary, and (10) improvement

suggestion. The meta‐codes are not mutually exclusive, and the

responses coded within them may overlap. Firstly, there were two

categories of less useful content for analysis. The meta‐code no

answer included answers which were not understandable and there-

fore not usable for the evaluation (n = 155). Respondents who did

not know how to answer the question were classified into the meta‐code no idea (n = 124), for instance: “I don't know at the moment as

these theses are absolutely new and I first have to process them”

(1354).

4.3.1 | Positive and thoughtful attitudes towardsthe guidelines

In summary, half of the participants’ answers (50%) were classified

into the meta‐code no criticism (n = 515). That means that they, in

total, agree with the guidelines and would not make many changes:

“I think the presented guideline is very coherent and complete”(241) or “The guideline is very well elaborated. No further sugges-

tions” (504). Many respondents are convinced of the guidelines: “Ireally like it. It should be short, give clear action hints and should

not be too instructive” (281). Another participant also has no criti-

cism, but he wants to be secured: “No criticism, it must be guaran-

teed that the mobile network is not dead in case of an emergency”(1801). However, some participants have doubts, for instance, on

the distribution of smartphones: “You think that everyone owns a

smartphone. What about the elder? People that reject smartphones

due to mistrust? Etc. I really like the presented suggestions, but the

question is what actually is realized in such situations” (177). Further

participants are sceptical about the application of guidelines: “I don'thave any criticism, but I don't think that all people will hold on to it”(262). In this context, it is also interesting that few respondents do

not have any improvement suggestions, but they think of people

KAUFHOLD ET AL. | 9

TABLE

5Results

ofch

i‐squ

ared

tables

forallqu

estions,w

ithco

rrelations

whe

reap

propriate

Q22:Before

Q23:During

Q24:After

Q25:Gen

eral

Gen

der

χ²(15,1024)=16.33,p=0.361;r=0.069

χ²(30,1024)=33.25,p=0.312;r=0.061

χ²(15,1024)=10.35,p=0.797;r=0.028

χ²(20,1024)=45.80,

p<

0.001;r=0.132

Age

χ²(75,1024)=65.99,p=0.762;r=0.002

χ²(150,1024)=127.10,p=0.913;r=0.016

χ²(75,1024)=74.33,p=0.500;r=0.012

χ²(100,1024)=126.26,

p<

0.05;r=0.149

Edu

cation

χ²(75,1024)=124.28,p<

0.001;r=−0.035

χ²(150,1024)=300.40,p<

0.001;r=0.052

χ²(75,1024)=136.64,p<

0.001;r=−0.007

χ²(100,1024)=150.50,

p=0.001;r=0.039

Inco

me

χ²(45,1024)=88.26,p<

0.001;r=0.030

χ²(90,1024)=89.21,p=0.504;r=0.042

χ²(45,1024)=50.71,p=0.259;r=0.036

χ²(60,1024)=79.69,

p<

0.05;r=0.066

Reg

ion

χ²(210,1024)=233.98,p=0.123

χ²(420,1024)=410.58,p=0.620

χ²(210,1024)=176.59,p=0.955

χ²(280,1024)=279.38,

p=0.449

Ove

rallsm

artpho

neusage

χ²(60,1024)=76.34,p=0.076;r=0.074

χ²(120,1024)=162.19,p<

0.01;r=0.083

χ²(60,1024)=106.78,p<

0.001;r=0.059

χ²(80,1024)=153.56,

p<

0.001;r=0.014

Ove

rallsocial

med

iausage

χ²(465,1024)=492.71,p=0.181;r=−0.030

χ²(930,1024)=951.63,p=0.304;r=0.000

χ²(465,1024)=449.38,p=0.690;r=0.015

χ²(620,1024)=528.815,

p=0.997;r=0.000

Postingin

social

med

iaχ²(60,1024)=96.49,p<

0.005;r=0.212

χ²(120,1024)=164.73,p<

0.005;r=0.151

χ²(60,1024)=100.98,p<

0.005;r=0.169

χ²(80,1024)=76.03,

p=0.605;r=0.028

Use

ofsocial

med

iain

crisis

situations

χ²(30,1024)=95.04,p<

0.001;r=0.252

χ²(60,1024)=100.76,p<

0.005;r=0.162

χ²(30,1024)=64.09,p<

0.001;r=0.186

χ²(40,1024)=65.06,

p<

0.01;r=0.010

Use

ofcrisis

apps

χ²(600,1024)=544.48,p=0.949;r=−0.004

χ²(1200,1024)=1122.61,p=0.945;r=−0.026

χ²(600,1024)=475.92,p=1.000;r=0.027

χ²(800,1024)=691.863,

p=0.998;r=0.000

Exp

ectations

ofCIto

usesocial

med

ia

χ²(60,1024)=241.73,p<

0.001;r=0.302

χ²(120,1024)=296.60,p<

0.001;r=0.282

χ²(60,1024)=261.72,p<

0.001;r=0.284

χ²(80,1024)=195.97,

p<

0.001;r=0.202

Non‐significant

resultsaredisplaye

dongrey

backgroun

d.

10 | KAUFHOLD ET AL.

who do not use social media: “I don't have any criticism, but there

are still many people who do not use social media – like me” (876).

4.3.2 | Suggestions for improvement on content,extent, and presentation

Furthermore, participants elaborated a variety of constructive im-

provement suggestions (n = 78). Although other categories also

include recommendations for improvement and thus responses may

overlap, we distinguished this particular meta‐code for a better over-

view of possible contributions to the improvement of our guidelines

and further research. For instance, participants suggested to add fur-

ther recommendations: “Only share correct and trustworthy informa-

tion, if so, with giving the source. If you detect misinformation, make

people aware of it” (70). Moreover, it is demanded that volunteers

coordinate each other and distribute auxiliary goods usefully. One

respondent also considers the question of time and has a suggestion:

“I ask myself if I have time to use Twitter or Facebook in case of an

emergency. I would like public terminals, which are readily accessible

and located in a safe place” (391). Besides Internet and social media,

one respondent makes clear that traditional media still should be

used for the dissemination of information (1615).

Also, there are some recommendations for the presentation of

the guideline (n = 26). For example, it should also be more under-

standable for children, foreign people, elder people, and disabled

persons. One participant recommends add some pictures: “As a text,

no one will really read it through” (697). In this context, the meta‐code extent matters as several respondents think the guideline is too

extensive (n = 36): “The guideline should be restricted to basic infor-

mation and ignore norms that are simply nice or desirable” (410).

One participant asks for a cheat sheet which summarizes the most

important aspects of the guideline to have a compact overview. On

the other hand, one interviewee even asks for more details: “Forinexperienced users of social media, a more explicit guideline would

be interesting, i.e., how I contact official entities via social media or

where I have to search on Facebook or Twitter” (1093).

4.3.3 | Scepticism about fake news and thenecessity of guidelines

As represented by the meta‐code scepticism (n = 52), several people

do not really trust the guideline as they have too many doubts con-

cerning the use, design, or privacy. For instance, one participant has

doubts about the usage type: “It should be emphasized VERY clearly

that social media do NOT replace the official numbers and channels”(157). Additionally, privacy is an important factor for few partici-

pants: “Private content of a conversation and chat reports should

not be investigated if there's no reason as we wouldn't have any pri-

vacy in society” (507). Furthermore, gazers were perceived as dis-

turbing factor in case of an emergency by a few participants (n = 3):

“If you don't want any gazers on the accident site, you shouldn't ask

for photos. I generally find it inconvenient to post something about

a catastrophe if the police don't ask for it. That quickly leads to the

distribution of misinformation as people have misperceptions in such

situations” (512).

Another meta‐code which shows the respondents’ doubts is the

category fake news (n = 26). As said before, people worry about the

dissemination of misinformation as there are much content and

many rumours on the Internet and social media: “I think, in case of

an emergency, social media are crap. You simply can create too

much fake news, and everyone thinks something happens” (463).

For this case, one participant suggests: “Actually none. Maybe Face-

book, Twitter and co. could develop bots which detect fake news or

evaluate user ratings (many dislikes) and automatically delete post-

ings” (1087). In summary, only nine people rate the guideline as un-

necessary (n = 9): “Why do we need it? Are there concrete reasons?

Such a guide can spread fear and feed speculations” (644). They are

very sceptical and do not believe that the guideline is very helpful in

case of an emergency: “It generally is totally unnecessary and a

waste of time as every normal human‐being, who thinks about it a

bit, will decide correctly in the concrete situation. And if not, a

‘guideline’ will not discourage him” (122).

5 | DISCUSSION AND CONCLUSION

In this paper, we outlined the risks of chaotic social media use

(Kaewkitipong et al., 2012; Perng et al., 2012; Valecha et al., 2013)

and presented guidelines for the use of social media in general and

in emergencies, revealing that most of them focus on the perspec-

tive of emergency services or organizations. Only three guidelines

considered the perspective of citizens before, during, and after

emergencies (Belfo et al., 2015; Emergency 2.0 Wiki, 2015; Helsloot

et al., 2015), and none of them was evaluated in terms of practical

value with a larger sample of citizens. That, firstly, led us to the

design of guidelines for citizens. Secondly, this paper contributes to

the qualitative and quantitative evaluation of the citizens’ guidelinesbased on a representative sample for Germany (N = 1024) stratified

for gender and age, but also ensuring a wide spread of the survey

sample in terms of region, education, and income. Considering our

research question RQ1: “How do citizens perceive guidelines for

social media use in emergencies?”, we have divided our results into

five separate research questions, which we will summarize in the fol-

lowing.

To what extent do participants agree to the presented guidelines

on social media behaviour before, during, and after emergencies

(RQ1.1)? Our results indicate that all proposed guidelines for social

media in crisis contexts achieve a high acceptance rate exceeding

50%. It is only slightly lower than 50% for anticipating the type of

information shared before a crisis and for helping and volunteering

in case an emergency occurs. The present guidelines are also rated

as helpful by over two‐thirds of respondents (72%).

How consistent is the agreement with guidelines across the cate-

gories, that is, concerning the use of social media before, during, and

after emergencies (RQ1.2)? Concerning the relationship between

responses to the individual guidelines, most participants tended to

agree to guidelines related to each other. In total, there is a

KAUFHOLD ET AL. | 11

significant and high correlation between agreements to the specific

categories. We can conclude that perception of the guideline impor-

tance is consistent across the stages (before, during, and after), while

responses to the general category, which show the highest agree-

ment, were not as highly correlated with the others (below

r = 0.500).

To what extent is the attitude towards the guidelines affected by

smartphone and social media usage in daily life and in emergencies

(RQ1.3)? The more often the individuals use their smartphone, post

in social media, or tend to use them in crisis situations, the more

they agree to our guidelines. Additionally, the agreement is consis-

tent with the expectation to read messages by CI providers in social

media during crises. However, overall social media or crisis app

usage does not have an effect on the responses. Therefore, guideli-

nes appeal most to active mobile social media users and those who

rely on them during emergencies.

How do demographic factors influence the agreement with

guidelines concerning the use of social media in emergencies

(RQ1.4)? Demographic factors have only a partial influence on the

responses, namely education and income, and to a small degree gen-

der and age. We can therefore conclude that they are generally rea-

sonable for most demographic groups, where age or economic

situation is not a strongly determining factor, although it might have

an influence on overall technology usage.

What are caveats and improvement suggestions towards the

guidelines on social media in emergencies (RQ1.5)? Most participants

had a positive reaction to the presented guidelines (see RQ1.1),

however, some improvements based on participants’ suggestions

had to be done to make the guidelines more relevant, concise,

appealing, and applicable for everyone. A reason for criticism, among

others, lies in the mode of presentation, which is often considered

too long and text based, as well as containing information that is not

vital in this context.

Our qualitative responses reflect the agreement previously

observed in RQ1.1: Half of the participants agree with the guidelines

as they are, while many also point out that these were not relevant

for people who do not use smartphones or social media, or if there

is no time or network for this in case of an emergency. These last

two comments are especially similar to results by Sutton, Woods,

and Vos (2017), who identified the time and effort invested into

information search as well as the technological access to reliable

information as hurdles to an effective crisis communication via social

media or websites. Therefore, it is suggested to expand the scope

and apply the guidelines to mass media as well. According to the

social‐mediated crisis communication model (Jin, Liu, & Austin,

2014), the media channel, along with information source and origin

of the crisis, is important for recipients’ perception of the crisis and

the affected organization. Generally, citizens tend to use personal

conversations and social media mostly for communication about a

crisis, but traditional media is still perceived as most trustworthy and

generates most positive attitudes (Austin, Liu, & Jin, 2012; Liu, Lin,

& Austin, 2013). In terms of our concrete recommendations, the

highest importance lies on reliability of sources and correctness

when reading and sharing information (Kaewkitipong et al., 2012), as

well as using social networks only as supplement to other media like

the emergency call. Another concern was fake news with the poten-

tial to cause panic and harm, and gazers that could be drawn to the

site of emergency.

The results of our study informed the final visualization of the

citizens’ guidelines (see Figure A1 in the Appendix 1). Firstly, the

content was reduced: The gathered feedback suggested to even

reduce the amount of text, that is, to omit general norms, and the

recommendations for social media use after emergencies received

the least consent of citizens. Secondly, some illustrations were used

to increase the overall presentation quality of the guideline.

5.1 | Limitations and future work

In future emergencies, the guidelines may be distributed in social

media, that is, Facebook and Twitter, to observe and analyse their

reception by citizens. Although the guidelines were evaluated with a

large sample representative for Germany, intercultural differences

could lead to different perceptions on the relevance of specific rec-

ommendations. For instance, prior research identified different risk

cultures across Europe, such as individual‐oriented, state‐oriented,and fatalistic risk cultures, which shape the behaviour and perception

of citizens (Dressel, 2015; Dressel & Pfeil, 2017). Despite disasters

are common (Höppe, 2015), the number of German citizens who

already experienced a crisis is relatively low (Reuter, Kaufhold, Leo-

pold, & Knipp, 2017). Therefore, this evaluation is based on a hypo-

thetical scenario in contrast to a case study, and the results could be

influenced by this fact. A specific case study could examine whether

participants who use social media already (partially) apply the strate-

gies addressed in the guidelines before, during, and after an emer-

gency, and how. Apart from an attitude towards the guidelines, we

could further analyse to what extent citizens would apply and

adhere to them in an actual crisis scenario.

Furthermore, guidelines for the use of social media for emer-

gency services were developed in our project. Despite the feed-

back of emergency managers during EAB workshops, no evaluation

was performed on the final guidelines. However, the specific needs

and perceptions of emergency managers in terms of citizens’ beha-viour and content creation in social media (Flizikowski, Hołubow-

icz, Stachowicz, Hokkanen, & Delavallade, 2014; Reuter et al.,

2016) are important precursors to optimize the interplay between

authorities and citizens during emergencies, and the design of both

guidelines.

ACKNOWLEDGEMENTS

The research project “EmerGent” was funded by a grant of the

European Union (FP7 No. 608352). The research group "KontiKat"

was funded by the German Federal Ministry of Education and

Research (BMBF No. 13N14351). We would like to thank all

members of our projects for their remarks and for distributing our

survey.

12 | KAUFHOLD ET AL.

REFERENCES

Albris, K. (2017). The switchboard mechanism: How social media con-

nected citizens during the 2013 floods in Dresden. Journal of Contin-

gencies and Crisis Management, 26, 350–357. https://doi.org/10.

1111/1468-5973.12201

Austin, L., Liu, B. F., & Jin, Y. (2012). How audiences seek out crisis infor-

mation: Exploring the social‐mediated crisis communication model.

Journal of Applied Communication Research, 40(2), 188–207. https://doi.org/10.1080/00909882.2012.654498

Baird, M. E. (2010). The Phases of Emergency Management. Retrieved

from http://www.memphis.edu/ifti/pdfs/cait_phases_of_emergency_

mngt.pdf

Belfo, J., Simão, L., Schmidt, S., Rhode, D., Freitag, S., Lück, A., … Villot,

E. (2015). Deliverable D2.271: iSAR+ Guidelines. Retrieved from:

http://www.cosmic-project.eu/sites/default/files/Deliverables_D6.1.2_

and_D6.2.2_Final_Guidelines_April_2015.pdf

Berliner Feuerwehr. (2012). Social‐Media‐Guideline – Empfehlungen für

einen sicheren Umgang mit sozialen Medien. Retrieved from: http://

www.berliner-feuerwehr.de/fileadmin/bfw/dokumente/Download/

2012/2012_01_SM-Guideline.pdf

Boin, A., Stern, E., & Sundelius, B. (2005). The Politics of Crisis Manage-

ment: Public Leadership Under Pressure. Cambridge: Cambridge Univer-

sity Press. https://doi.org/10.1017/CBO9780511490880

Brynielsson, J., Granåsen, M., Lindquist, S., Narganes Quijano, M., Nilsson,

S., & Trnka, J. (2017). Informing crisis alerts using social media: Best

practices and proof of concept. Journal of Contingencies and Crisis

Management (JCCM), 26, 28–40. https://doi.org/10.1111/1468-5973.12195

Bundesanstalt Technisches Hilfswerk. (2011). Verhalten in sozialen Net-

zwerken, 53127.

Caragea, C., Squicciarini, A., Stehle, S., Neppalli, K., & Tapia, A. (2014).

Mapping Moods: Geo‐Mapped Sentiment Analysis During Hurricane

Sandy. In S. R. Hiltz, M. S. Pfaff, L. Plotnick & A. C. Robinson (Eds.),

Proceedings of the 11th International Conference on Information Sys-

tems for Crisis Response and Management (ISCRAM) (pp. 642–651).Brussels, Belgium: ISCRAM.

Daimler, A. G. (2012). Social Media Leitfaden – 10 Tipps zum Umgang

mit Social Media. Retrieved from: https://www.daimler.com/dokume

nte/konzern/sonstiges/daimler-socialmedialeitfaden-de.pdf

Deutscher Caritasverband. (2014). Das Soziale ins Netz bringen – die

Caritas und soziale Medien. Retrieved from: https://www.caritas.de/

cms/contents/caritasde/medien/dokumente/zentraledokumente/soc

ialmedialeitlinie/social_media_leitlinien_caritas_mitarbeiter.pdf

Deutsches Rotes Kreuz. (2012). Ein Leitfaden zum Umgang mit Social

Media im DRK. Retrieved from: https://www.drk-baden-wuerttembe

rg.de/fileadmin/Eigene_Bilder_und_Videos/Publikationen/Mediathek/

Social_Media_Leitfaden_DRK.pdf

Dressel, K. (2015). Risk culture and crisis communication. International

Journal of Risk Assessment and Management, 18(2), 115. https://doi.

org/10.1504/IJRAM.2015.069020

Dressel, K., & Pfeil, P. (2017). Socio‐cultural factors of risk and crisis

communication: Crisis communication or what civil protection agen-

cies should be aware of when communicating with the public in crisis

situations. In M. Klafft (Ed.), Risk and Crisis Communication for Disaster

Prevention and Management (pp. 64–76). Wilhelmshaven: epubli.

Retrieved from https://www.epubli.de/shop/buch/Risk-and-Crisis-

Communication-for-Disaster-Prevention-and-Management-Michael-

Klafft-Michael-Klafft-9783745054484/62951#beschreibung

Emergency 2.0 Wiki. (2015). Emergency 2.0 Wiki. Retrieved from http://

emergency20wiki.org/

Eriksson, M., & Olsson, E. K. (2016). Facebook and Twitter in crisis com-

munication: A comparative study of crisis communication profession-

als and citizens. Journal of Contingencies and Crisis Management, 24(4),

198–208. https://doi.org/10.1111/1468-5973.12116

Flizikowski, A., Hołubowicz, W., Stachowicz, A., Hokkanen, L., & Delaval-

lade, T. (2014). Social Media in Crisis Management – the iSAR + Pro-

ject Survey. In S. R. Hiltz, M. S. Pfaff, L. Plotnick & P. C. Shih (Eds.),

Proceedings of the 11th International Conference on Information Sys-

tems for Crisis Response and Management (ISCRAM) (pp. 707–711).Brussels, Belgium: ISCRAM.

German Federal Agency for Civic Education. (2016). Datenreport 2016.

Ein Sozialbericht für die Bundesrepublik Deutschland.

Gizikis, A., O'Brien, T., Susaeta, I. G., Moi, M., Schubert, A., Reuter, C., & ,

…, U. (2017). EmerGent Deliverable 7.3: Guidelines to increase the

benefit of social media in emergencies. Paderborn. Retrieved from

http://www.fp7-emergent.eu/publications/20170529_D7.3_Guideline

s_to_increase_the_benefit_of_social_media_EmerGent.pdf

Gizikis, A., Susaeta, I. G., Spielhofer, T., & Bizjak, G. (2017). EmerGent

Deliverable 2.7: Workshop III. Paderborn. Retrieved from http://

www.fp7-emergent.eu/wp-content/uploads/2017/04/20170330_D2.

7_Workshop_III_EmerGent_pub.pdf

Helsloot, I., & Ruitenberg, A. (2004). Citizen response to disasters: A sur-

vey of literature and some practical implications. Journal of Contingen-

cies and Crisis Management, 12(3), 98–111. https://doi.org/10.1111/j.0966-0879.2004.00440.x

Helsloot, I., deVries, D., Groenendaal, J., Scholtens, A., in ‘t Veld, M., van-

Melick, G., … Vontas, A. M. (2015). Deliverable D6.1 & D6.2: Guide-

lines for the use of new media in crisis situations; Project “Cosmic –The COntribution of Social Media In Crisis management”. https://doi.org/10.5281/zenodo.16235

Hiltz, S. R., & Plotnick, L. (2013). Dealing with Information Overload

When Using Social Media for Emergency Management: Emerging

Solutions. In T. Comes, F. Fiedrich, S. Fortier, J. Geldermann & T.

Müller (Eds.), Proceedings of the 10th International Conference on Infor-

mation Systems for Crisis Response and Management (ISCRAM) (pp.

823–827). Brussels, Belgium: ISCRAM.

Homeland Security. (2013). Lessons Learned: Social Media and Hurricane

Sandy ‐ Virtual Social Media Working Group and Group DHS First

Responders.

Höppe, P. (2015). Naturkatastrophen – immer häufiger, heftiger, tödli-

cher, teurer? Retrieved from http://www.munichre-foundation.org/

dms/MRS/Documents/20150303_DF2015_Hoeppe_Disasters/

DF2015_March_Handout_Hoeppe.pdf

IBM. (2014). Statistical Package for the Social Sciences. Retrieved from

https://www.ibm.com/products/spss-statistics

Jin, Y., Liu, B. F., & Austin, L. L. (2014). Examining the role of social

media in effective crisis management. Communication Research, 41(1),

74–94. https://doi.org/10.1177/0093650211423918Jurgens, M., & Helsloot, I. (2017). The effect of social media on the

dynamics of (self) resilience during disasters: A literature review. Jour-

nal of Contingencies and Crisis Management (JCCM), 26, 79–88.https://doi.org/10.1111/1468-5973.12212

Kaewkitipong, L., Chen, C., & Ractham, P. (2012). Lessons Learned from

the Use of Social Media in Combating a Crisis: A Case Study of 2011

Thailand Flooding Disaster. In Proceedings of the Thirty Third Interna-

tional Conference on Information Systems (ICIS) (pp. 1–17). Atlanta, GA:AISeL.

Kaufhold, M.‐A., & Reuter, C. (2016). The self‐organization of digital vol-

unteers across social media: The case of the 2013 European floods

in Germany. Journal of Homeland Security and Emergency Manage-

ment (JHSEM), 13(1), 137–166. https://doi.org/10.1515/jhsem-2015-

0063

Kogan, M., Palen, L., & Anderson, K. M. (2015). Think Local, Retweet

Global: Retweeting by the Geographically‐Vulnerable during Hurri-

cane Sandy. In Proceedings of the 18th ACM Conference on Computer

Supported Cooperative Work & Social Computing (CSCW ‘15) (pp. 981–993). New York, NY: ACM.

Li, J., & Rao, H. R. (2010). Twitter as a rapid response news service: An

exploration in the context of the 2008 China earthquake. The

KAUFHOLD ET AL. | 13

Electronic Journal on Information Systems in Developing Countries, 42,

1–22. https://doi.org/10.1002/(ISSN)1681-4835

Liu, B. F., Lin, Y., & Austin, L. L. (2013). The tendency to tell: Understand-

ing publics’ communicative responses to crisis information form and

source. Journal of Public Relations Research, 25, 51–67. https://doi.

org/10.1080/1062726X.2013.739101

Mirbabaie, M., & Zapatka, E. (2017). Sensemaking in Social Media Crisis

Communication – A Case Study on the Brussels Bombings in 2016.

In Twenty‐Fifth European Conference on Information Systems (ECIS) (pp.

2169–2186). Atlanta, GA: AISeL.O'Brien, T., Gizikis, A., Brugghemans, B., Spielhofer, T., Moi, M., & Friberg, T.

(2016). EmerGent Deliverable 2.6:Workshops I and II. Paderborn.

Palen, L., & Hughes, A. L. (2018). Social Media in Disaster Communica-

tion. In H. Rodríguez, W. Donner, & J. E. Trainor (Eds.), Handbook of

Disaster Research (pp. 497–518). Cham: Springer International Pub-

lishing. https://doi.org/10.1007/978-3-319-63254-4_24

Perng, S.‐Y., Büscher, M., Wood, L., Halvorsrud, R., Stiso, M., Ramirez, L.,

& Al‐Akkad, A. (2012). Peripheral response: Microblogging during the

22/7/2011 Norway attacks. In L. Rothkrantz, J. Ristvej & Z. Franco

(Eds.), Proceedings of the 9th International Conference on Information

Systems for Crisis Response and Management (ISCRAM) (pp. 1–11).Brussels, Belgium: ISCRAM.

Plotnick, L., & Hiltz, S. R. (2016). Barriers to use of social media by emer-

gency managers. Journal of Homeland Security and Emergency Manage-

ment, 13(2), https://doi.org/10.1515/jhsem-2015-0068

Qu, Y., Huang, C., Zhang, P., & Zhang, J. (2011). Microblogging after a

Major Disaster in China: A Case Study of the 2010 Yushu Earthquake.

In Proceedings of the ACM 2011 conference on Computer supported

cooperative work (CSCW ‘11) (pp. 25–34). New York, NY: ACM.

Qu, Y., Wu, P. F., & Wang, X. (2009). Online Community Response to

Major Disaster: A Study of Tianya Forum in the 2008 Sichuan Earth-

quake. In Proceedings of the Hawaii International Conference on System

Sciences (HICSS). Waikoloa, USA: IEEE.

Reuter, C., Heger, O., & Pipek, V. (2013). Combining Real and Virtual Vol-

unteers through Social Media. In T. Comes, F. Fiedrich, S. Fortier, J.

Geldermann & T. Müller (Eds.), Proceedings of the 10th International

Conference on Information Systems for Crisis Response and Management

(ISCRAM) (pp. 780–790). New York, NY: ACM. https://doi.org/10.

1126/science.1060143

Reuter, C., & Kaufhold, M.‐A. (2018). Fifteen years of social media in

emergencies: A retrospective review and future directions for crisis

informatics. Journal of Contingencies and Crisis Management (JCCM),

26, 41–57. https://doi.org/10.1111/1468-5973.12196Reuter, C., Kaufhold, M.‐A., Leopold, I., & Knipp, H. (2017). KATWARN,

NINA or FEMA? Multi‐Method Study on Distribution, Use and Public

Views on Crisis Apps. In Twenty‐Fifth European Conference on Infor-

mation Systems (ECIS) (pp. 2187–2201). Atlanta, GA: AISeL.Reuter, C., Ludwig, T., Kaufhold, M.‐A., & Spielhofer, T. (2016). Emer-

gency services attitudes towards social media: A quantitative and

qualitative survey across Europe. International Journal on Human‐Com-

puter Studies (IJHCS), 95, 96–111. https://doi.org/10.1016/j.ijhcs.

2016.03.005

Reuter, C., & Schröter, J. (2015). Microblogging during the European floods

2013: What Twitter may contribute in German emergencies. Interna-

tional Journal of Information Systems for Crisis Response and Management

(IJISCRAM), 7(1), 22–41. https://doi.org/10.4018/IJISCRAMReuter, C., & Spielhofer, T. (2017). Towards social resilience: A quantitative

and qualitative survey on citizens’ perception of social media in emer-

gencies in Europe. Technological Forecasting and Social Change (TFSC),

121, 168–180. https://doi.org/10.1016/j.techfore.2016.07.038Schmidt, A., Wolbers, J., Ferguson, J., & Boersma, K. (2017). Are you

Ready2Help? Conceptualizing the management of online and onsite

volunteer convergence. Journal of Contingencies and Crisis Manage-

ment, 26, 338–349. https://doi.org/10.1111/1468-5973.12200Statista. (2016). Bevölkerung Deutschlands nach Altersgruppen 2016.

Retrieved from https://de.statista.com/statistik/daten/studie/1365/

umfrage/bevoelkerung-deutschlands-nach-altersgruppen/

Statistisches Bundesamt (Destatis). (2016). Bildungsstand der Bevölker-

ung. Retrieved from https://www.destatis.de/DE/Publikationen/The

matisch/BildungForschungKultur/Bildungsstand/BildungsstandBevoe

lkerung5210002167004.pdf?__blob=publicationFile

Stieglitz, S., Mirbabaie, M., Ross, B., & Neuberger, C. (2018). Social media

analytics – Challenges in topic discovery, data collection, and data

preparation. International Journal of Information Management, 39,

156–168. https://doi.org/10.1016/j.ijinfomgt.2017.12.002

Strauss, A. L. (1987). Qualitative Analysis for Social Scientists. Cambridge:

Cambridge Press. https://doi.org/10.1017/CBO9780511557842

Sutton, J., Woods, C., & Vos, S. C. (2017). Willingness to click: Risk infor-

mation seeking during imminent threats. Journal of Contingencies and

Crisis Management (JCCM), 26, 283–294. 1–12https://doi.org/10.1111/1468-5973.12197

Valecha, R., Oh, O., & Rao, R. (2013). An Exploration of Collaboration

over Time in Collective Crisis Response during the Haiti 2010 Earth-

quake. In R. Baskerville & M. Chau (Eds.), Proceedings of the Thirty

Fourth International Conference on Information Systems (ICIS) (pp. 1–10). Atlanta, GA: AISeL.

How to cite this article: Kaufhold M‐A, Gizikis A, Reuter C,

Habdank M, Grinko M. Avoiding chaotic use of social media

before, during, and after emergencies: Design and evaluation

of citizens’ guidelines. J Contingencies and Crisis Management.

2018;00:1–16. https://doi.org/10.1111/1468-5973.12249

14 | KAUFHOLD ET AL.

APPENDIX

VISUALIZATION OF THE GUIDELINES

F IGURE A1 Visualization of the Guidelines

KAUFHOLD ET AL. | 15

TABLE

A1

Correlationco

efficien

ts(Spe

arman

'sRho

)be

twee

nresponses

forallgu

idelines

(p<

0.001)

Q22.1

Q22.2

Q22.3

Q22.4

Q23.1

Q23.2

Q23.3

Q23.4

Q23.5

Q23.6

Q23.7

Q23.8

Q23.9

Q24.1

Q24.2

Q24.3

Q24.4

Q25.1

Q25.2

Q25.3

Q25.4

Q25.5

Q25.6

Q25.6

0.344

0.131

0.270

0.438

0.335

0.529

0.358

0.573

0.582

0.560

0.441

0.215

0.318

0.303

0.281

0.378

0.391

0.622

0.668

0.606

0.574

0.785

Q25.6

Q25.5

0.373

0.160

0.297

0.480

0.368

0.545

0.376

0.597

0.609

0.564

0.464

0.234

0.366

0.321

0.310

0.387

0.409

0.660

0.713

0.653

0.673

0.785

Q25.5

Q25.4

0.329

0.185

0.268

0.376

0.311

0.423

0.370

0.448

0.496

0.432

0.392

0.289

0.371

0.306

0.275

0.372

0.382

0.528

0.605

0.625

0.673

0.574

Q25.4

Q25.3

0.324

0.192

0.254

0.393

0.333

0.460

0.324

0.478

0.478

0.452

0.402

0.211

0.302

0.265

0.288

0.348

0.362

0.613

0.695

0.625

0.653

0.606

Q25.3

Q25.2

0.367

0.160

0.288

0.462

0.388

0.545

0.367

0.592

0.597

0.545

0.441

0.233

0.390

0.326

0.305

0.381

0.400

0.740

0.695

0.605

0.713

0.668

Q25.2

Q25.1

0.388

0.169

0.300

0.433

0.412

0.510

0.360

0.558

0.538

0.512

0.439

0.229

0.383

0.355

0.312

0.387

0.383

0.740

0.613

0.528

0.660

0.622

Q25.1

Q24.4

0.434

0.339

0.393

0.453

0.428

0.356

0.431

0.383

0.406

0.432

0.501

0.439

0.480

0.498

0.531

0.718

0.383

0.400

0.362

0.382

0.409

0.391

Q24.4

Q24.3

0.431

0.373

0.401

0.444

0.447

0.319

0.432

0.374

0.400

0.396

0.500

0.440

0.461

0.520

0.557

0.718

0.387

0.381

0.348

0.372

0.387

0.378

Q24.3

Q24.2

0.412

0.379

0.401

0.400

0.422

0.268

0.399

0.316

0.310

0.390

0.452

0.434

0.435

0.533

0.557

0.531

0.312

0.305

0.288

0.275

0.310

0.281

Q24.2

Q24.1

0.568

0.494

0.486

0.461

0.587

0.295

0.495

0.362

0.343

0.373

0.492

0.539

0.531

0.533

0.520

0.498

0.355

0.326

0.295

0.306

0.321

0.303

Q24.1

Q23.9

0.421

0.413

0.411

0.450

0.490

0.363

0.497

0.437

0.435

0.452

0.520

0.658

0.531

0.435

0.461

0.480

0.383

0.390

0.302

0.371

0.366

0.318

Q23.9

Q23.8

0.394

0.486

0.426

0.370

0.422

0.237

0.503

0.289

0.294

0.342

0.510

0.658

0.539

0.434

0.440

0.439

0.229

0.233

0.211

0.289

0.234

0.215

Q23.8

Q23.7

0.472

0.332

0.421

0.473

0.495

0.468

0.531

0.555

0.579

0.596

0.510

0.520

0.492

0.542

0.500

0.501

0.439

0.441

0.402

0.392

0.464

0.441

Q23.7

Q23.6

0.392

0.209

0.304

0.440

0.411

0.570

0.453

0.689

0.672

0.596

0.342

0.452

0.373

0.390

0.396

0.432

0.512

0.545

0.452

0.432

0.564

0.560

Q23.6

Q23.5

0.397

0.201

0.309

0.480

0.418

0.608

0.483

0.778

0.672

0.579

0.294

0.435

0.343

0.310

0.400

0.406

0.538

0.597

0.478

0.496

0.609

0.582

Q23.5

Q23.4

0.390

0.209

0.334

0.495

0.419

0.630

0.488

0.778

0.689

0.555

0.289

0.437

0.362

0.316

0.374

0.383

0.558

0.592

0.478

0.448

0.597

0.573

Q23.4

Q23.3

0.440

0.425

0.431

0.462

0.532

0.469

0.488

0.483

0.453

0.531

0.503

0.497

0.495

0.399

0.432

0.431

0.360

0.367

0.324

0.370

0.376

0.358

Q23.3

Q23.2

0.395

0.183

0.282

0.471

0.428

0.469

0.630

0.608

0.570

0.468

0.237

0.363

0.295

0.268

0.319

0.356

0.510

0.545

0.460

0.423

0.545

0.529

Q23.2

Q23.1

0.614

0.451

0.494

0.534

0.428

0.532

0.419

0.418

0.411

0.495

0.422

0.490

0.587

0.422

0.447

0.428

0.412

0.388

0.333

0.311

0.368

0.335

Q23.1

Q22.4

0.593

0.442

0.610

0.534

0.471

0.462

0.495

0.480

0.440

0.473

0.370

0.450

0.461

0.400

0.444

0.453

0.433

0.462

0.393

0.376

0.480

0.438

Q22.4

Q22.3

0.539

0.564

0.610

0.494

0.282

0.431

0.334

0.309

0.304

0.421

0.426

0.411

0.486

0.401

0.401

0.393

0.300

0.288

0.254

0.268

0.297

0.270

Q22.3

Q22.2

0.545

0.564

0.442

0.452

0.183

0.425

0.209

0.201

0.209

0.332

0.486

0.413

0.494

0.379

0.373

0.339

0.169

0.160

0.192

0.185

0.160

0.131

Q22.2

Q22.1

0.545

0.539

0.593

0.614

0.395

0.440

0.390

0.397

0.392

0.472

0.394

0.421

0.568

0.412

0.431

0.434

0.388

0.367

0.324

0.329

0.373

0.344

Q22.1

Q22.1

Q22.2

Q22.3

Q22.4

Q23.1

Q23.2

Q23.3

Q23.4

Q23.5

Q23.6

Q23.7

Q23.8

Q23.9

Q24.1

Q24.2

Q24.3

Q24.4

Q25.1

Q25.2

Q25.3

Q25.4

Q25.5

Q25.6

APPENDIX

CORRELATIO

NSBETW

EEN

RESPONSESFOR

ALLGUID

ELIN

ES

16 | KAUFHOLD ET AL.